Flood Risk Management: Research
and Practice
Editors
Paul Samuels
Water Management, HR Wallingford, Wallingford, Oxfordshire, UK
Stephen Huntington
HR Wallingford Group, Wallingford, Oxfordshire, UK
William Allsop
Coastal Structures, HR Wallingford, Wallingford, Oxfordshire, UK
Jackie Harrop
CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business © 2009 Taylor & Francis Group, London, UK
Improving the understanding of the risk from groundwater flooding in the UK by D.M.J. Macdonald, J.P. Bloomfield, A.G. Hughes, A.M. MacDonald, B. Adams & A.A. McKenzie
© British Geological Survey
The worst North Sea storm surge for 50 years: Performance of the forecasting system and implications for decision makers by K.J. Horsburgh, J. Williams, J. Flowerdew, K. Mylne, S. Wortley
© Crown copyright
Typeset by Vikatan Publishing Solutions (P) Ltd., Chennai, India.
Printed and bound in Great Britain by Antony Rowe (A CPI-group Company), Chippenham, Wiltshire. All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher.
Although all care is taken to ensure integrity and the quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to the property or persons as a result of operation or use of this publication and/or the information contained herein.
Published by: CRC Press/Balkema
P.O. Box 447, 2300 AK Leiden, The Netherlands e-mail: Pub.NL@taylorandfrancis.com
www.crcpress.com – www.taylorandfrancis.co.uk – www.balkema.nl
Table of contents
Foreword
XIX
Committees
XXI
KEYNOTE PRESENTATION
Coastal flooding: A view from a practical Dutchman on present and future strategies 3 J.W. van der Meer
TECHNICAL PRESENTATIONS
Inundation modelling
Recent development and application of a rapid flood spreading method 15 J. Lhomme, P. Sayers, B. Gouldby, P. Samuels, M. Wills & J. Mulet-Marti
Hydrodynamic modelling and risk analysis in RAMFLOOD project 16
E. Bladé, M. Gómez-Valentín, J. Dolz, M. Sánchez-Juny, J. Piazzese, E. Oñate & G. Corestein
Testing and application of a practical new 2D hydrodynamic model 17
J. Gutierrez Andres, J. Lhomme, A. Weisgerber, A. Cooper, B. Gouldby & J. Mulet-Marti
Floods study through coupled numerical modeling of 2D surface and sewage network flows 18 C. Coulet, L. Evaux & A. Rebaï
Modelling of flooding and analysis of pluvial flood risk – demo case of UK catchment 19 J.P. Leitão, S. Boonya-aroonnet, Cˇ. Maksimovic´, R. Allitt & D. Prodanovic´
An integrated approach to modelling surface water flood risk in urban areas 21 J.B. Butler, D.M. Martin, E.M. Stephens & L. Smith
Estimation of flood inundation probabilities using global hazard indexes based
on hydrodynamic variables 22
G.T. Aronica, P. Fabio, A. Candela & M. Santoro
Flood modeling for risk evaluation – a MIKE FLOOD vs. SOBEK 1D2D benchmark study 23 P. Vanderkimpen, E. Melger & P. Peeters
Comparing forecast skill of inundation models of differing complexity: The case of
Upton upon Severn 25
K. Srinivas, M. Werner & N. Wright
Comparison of varying complexity numerical models for the prediction of flood inundation
in Greenwich, UK 26
VI
Grid resolution dependency in inundation modelling: A case study 29
S. Néelz & G. Pender
2D overland flow modelling using fine scale DEM with manageable runtimes 30 J.N. Hartnack, H.G. Enggrob & M. Rungø
Detailed 2D flow simulations as an onset for evaluating socio-economic impacts of floods 31 B.J. Dewals, S. Detrembleur, P. Archambeau, S. Erpicum & M. Pirotton
Ensemble Prediction of Inundation Risk and Uncertainty arising from Scour
(EPIRUS): An overview 33
Q. Zou, D. Reeve, I. Cluckie, S. Pan, M.A. Rico-Ramirez, D. Han, X. Lv, A. Pedrozo-Acuña & Y. Chen
Flood risk assessment using broad scale two-dimensional hydraulic modelling – a case study
from Penrith, Australia 35
H. Rehman, R. Thomson & R. Thilliyar
Modelling and analysis of river flood impacts on sewage networks in urban areas 36 A. Kron, P. Oberle, A. Wetzel & N. Ettrich
Coastal flood risk modelling in a data rich world 38
R.D. Williams, M.R. Lawless & J. Walker
A multi-scale modelling procedure to quantify effects of upland land management on flood risk 40 H.S. Wheater, B.M. Jackson, O. Francis, N. McIntyre, M. Marshall, I. Solloway,
Z. Frogbrook & B. Reynolds
Updating flood maps using 2D models in Italy: A case study 41
F. Nardi, J.S. O’Brien, G. Cuomo, R. Garcia & S. Grimaldi
Real-time validation of a digital flood-inundation model: A case-study from Lakes Entrance,
Victoria, Australia 43
P.J. Wheeler, J. Kunapo, M.L.F. Coller, J.A. Peterson & M. McMahon
Dispelling the myths of urban flood inundation modelling 44
D. Fortune
Flood risk in urban areas caused by levee breaching 45
A. Paquier, C. Peyre, N. Taillefer & M. Chenaf
RISK-EOS flood risk analysis service for Europe 46
V. Holzhauer, M. Müller & A. Assmann
Flood inundation modelling: Model choice and application 47
N. Asselman, J. ter Maat, A. de Wit, G. Verhoeven, S. Soares Frazão, M. Velickovic, L. Goutiere, Y. Zech, T. Fewtrell & P. Bates
Risk maps of torrential rainstorms 48
A. Assmann, M. Krischke & E. Höppner
Decision Support System for flood forecasting and risk mitigation in the context of
Romanian water sector 49
I. Popescu, A. Jonoski & A. Lobbrecht
Developing a rapid mapping and monitoring service for flood management using remote
sensing techniques 50
V. Craciunescu, C. Flueraru, G. Stancalie & A. Irimescu
A framework for Decision Support Systems for flood event management – application to the
Thames and the Schelde Estuaries 51
J.C. Neal, P.D. Bates, T.J. Fewtrell, N.G. Wright, I. Villanueva, N.M. Hunter & M.S. Horritt
Experience of 1D and 2D flood modelling in Australia – a guide to model selection based
on channel and floodplain characteristics 55
J.M. Hannan & J. Kandasamy
Computationally efficient flood water level prediction (with uncertainty) 56 K. Beven, P. Young, D. Leedal & R. Romanowicz
Optimization of 2D flood models by semi-automated incorporation of flood diverting
landscape elements 57
P. Vanderkimpen, P. Peeters & K. Van der Biest
Understanding the runoff response of the Ourthe catchment using spatial and temporal
characteristics of the storm field obtained by radar 59
P. Hazenberg, H. Leijnse, R. Uijlenhoet & L. Delobbe
The importance of spill conceptualizations and head loss coefficients in a quasi two-dimensional
approach for river inundation modelling 60
M.F. Villazón & P. Willems
Inundation scenario development for damage evaluation in polder areas 61 L.M. Bouwer, P. Bubeck, A.J. Wagtendonk & J.C.J.H. Aerts
System analysis
Importance of river system behaviour in assessing flood risk 65
M.C.L.M. van Mierlo, T. Schweckendiek & W.M.G. Courage
Development and evaluation of an integrated hydrological modelling tool for the Water Framework
Directive and Floods Directive 66
M.B. Butts, E. Fontenot, M. Cavalli, C.Y. Pin, T.S. Jensen, T. Clausen & A. Taylor
A comparison of modelling methods for urban flood risk assessment 67
C.J. Digman, T. Bamford, D.J. Balmforth, N.M. Hunter & S.G. Waller
Coastal flood risk analysis driven by climatic and coastal morphological modelling 68 M.J. Walkden, J.W. Hall, R. Dawson, N. Roche & M. Dickson
Micro-scale analysis of flood risk at the German Bight Coast 69
G. Kaiser, S.D. Hofmann, H. Sterr & A. Kortenhaus
Flood hazard mapping for coastal storms in the Delta Ebro 70
D. Alvarado-Aguilar & J.A. Jiménez
RAMWASS Decision Support System (DSS) for the risk assessment of water-sediment-soil systems – application of a DSS prototype to a test site
in the lower part of the Elbe river valley, Germany 71
B. Koppe, B. Llacay & G. Peffer
Radar based nowcasting of rainfall events – analysis and assessment
VIII
On the quality of Pareto calibration solutions of conceptual rainfall-runoff models 73 A.-R. Nazemi, A.H. Chan, A. Pryke & X. Yao
Model reuse and management in flood risk modelling 74
R. Khatibi
International programmes
Flood Risk from Extreme Events (FREE): A NERC-directed research programme – understanding
the science of flooding 77
C.G. Collier
Advances in flood risk management from the FLOODsite project 78
P.G. Samuels, M.W. Morris, P. Sayers, J-D. Creutin, A. Kortenhaus, F. Klijn, E. Mosselman, A. van Os & J. Schanze
The Tyndall Centre Coastal Simulator and Interface (CoastS) 79
R.J. Nicholls, M. Mokrech, S.E. Hanson, P. Stansby, N. Chini, M. Walkden, R. Dawson, N. Roche, J.W. Hall, S.A. Nicholson-Cole, A.R. Watkinson, S.R. Jude, J.A. Lowe, J. Leake, J. Wolf, C. Fontaine, M. Rounsvell & L. Acosta-Michlik
The social impacts of flooding in Scotland: A national and local analysis 81 A. Werritty, D.M. Houston, M. Jobe, T. Ball, A.C.W. Tavendale & A.R. Black
The Flood Risk Management Research Consortium (FRMRC) 82
I.D. Cluckie
EIB financing for flood risk mitigation 83
C. Gleitsmann
One nation, one policy, one program flood risk management 84
P.D. Rabbon, L.J. Zepp & J.R. Olsen
Toward a transnational perspective on flood-related research in Europe – experiences from
the CRUE ERA-Net 86
A. Pichler, V. Jackson, S. Catovsky & T. Deppe
Infrastructure and assets
Hazards from wave overtopping 89
W. Allsop, T. Bruce, T. Pullen & J. van der Meer
Time-dependent reliability analysis of anchored sheet pile walls 91
F.A. Buijs, P.B. Sayers, J.W. Hall & P.H.A.J.M. van Gelder
Analysis of tsunami hazards by modelling tsunami wave effects 93
T. Rossetto, W. Allsop, D. Robinson, I. Chavet & P.-H. Bazin
Influence of management and maintenance on erosive impact of wave overtopping on grass
covered slopes of dikes; Tests 95
G.J. Steendam, W. de Vries, J.W. van der Meer, A. van Hoven, G. de Raat & J.Y. Frissel Sea wall or sea front? Looking at engineering for Flood and Coastal Erosion Risk
Management through different eyes 97
J. Simm
The new turner contemporary gallery – an example of an urban coastal
flood risk assessment 98
Calculation of fragility curves for flood defence assets 102 J.W. van der Meer, W.L.A. ter Horst & E.H. van Velzen
Reservoir flood risk in the UK 104
A.L. Warren
Modelling breach initiation and growth 105
M.W. Morris, M.A.A.M. Hassan, A. Kortenhaus, P. Geisenhainer, P.J. Visser & Y. Zhu
A probabilistic failure model for large embankment dams 107
N.P. Huber, J. Köngeter & H. Schüttrumpf
Reliability analysis of flood defence structures and systems in Europe 109 P. van Gelder, F. Buijs, W. ter Horst, W. Kanning, C. Mai Van, M. Rajabalinejad, E. de Boer, S. Gupta,
R. Shams, N. van Erp, B. Gouldby, G. Kingston, P. Sayers, M. Wills, A. Kortenhaus & H.-J. Lambrecht
PCRIVER—software for probability based flood protection 111
U. Merkel, B. Westrich & A. Moellmann
Representing fragility of flood and coastal defences: Getting into the detail 112 J. Simm, B. Gouldby, P. Sayers, J-J. Flikweert, S. Wersching & M. Bramley
Application of 3D serious games in levee inspection education 113
M. Hounjet, J. Maccabiani, R. van den Bergh & C. Harteveld
Strategic appraisal of flood risk management options over extended timescales: Combining
scenario analysis with optimization 114
J.W. Hall, T.R. Phillips, R.J. Dawson, S.L. Barr, A.C. Ford, M. Batty, A. Dagoumas & P.B. Sayers Embedding new science into practice – lessons from the development and application
of a Performance-based asset management system 116
C. Mitchell, O. Tarrant, D. Denness, P. Sayers, J. Simm & M. Bramley
Study of flood embankment behaviour induced by air entrapment 117
D. Lesniewska, H. Zaradny, P. Bogacz & J. Kaczmarek
Assessment of flood retention in polders using an interlinked one-two-dimensional hydraulic model 119 M. Kufeld, H. Schüttrumpf & D. Bachmann
Fragility curve calculation for technical flood protection measures by the Monte Carlo analysis 120 D. Bachmann, N.P. Huber & H. Schüttrumpf
Application of GMS system in the Czech Republic – practical use of IMPACT, FLOODSite
and GEMSTONE projects outcomes 121
Z. Boukalová & V. Beneš
Failure modes and mechanisms for flood defence structures 122
M.W. Morris, W. Allsop, F.A. Buijs, A. Kortenhaus, N. Doorn & D. Lesniewska
Non-structural approaches (CRUE project)
Flood risk map perception through experimental graphic semiology 127
X
Efficiency of non-structural flood mitigation measures: “room for the river”
and “retaining water in the landscape” 130
S. Salazar, F. Francés, J. Komma, G. Blöschl, T. Blume, T. Francke & A. Bronstert
Flood risk reduction by PReserving and restOring river FLOODPLAINs – PRO_FLOODPLAIN 131 H. Habersack, C. Hauer, B. Schober, E. Dister, I. Quick, O. Harms, M. Wintz,
E. Piquette & U. Schwarz
The use of non structural measures for reducing the flood risk in small urban catchments 132 E. Pasche, N. Manojlovic, D. Schertzer, J.F. Deroubaix, I. Tchguirinskaia, E. El Tabach, R. Ashley,
R. Newman, I. Douglas, N. Lawson & S. Garvin
EWASE—Early Warning Systems Efficiency: Evaluation of flood forecast reliability 134 K. Schröter, M. Ostrowski, M. Gocht, B. Kahl, H.-P. Nachtnebel, C. Corral & D. Sempere-Torres
Flood risk assessment in an Austrian municipality comprising the evaluation of effectiveness
and efficiency of flood mitigation measures 135
C. Neuhold & H.-P. Nachtnebel
EWASE—Early Warning Systems Efficiency – risk assessment and efficiency analysis 136 M. Gocht, K. Schröter, M. Ostrowski, C. Rubin & H.P. Nachtnebel
Flood risk management strategies in European Member States considering structural and
non-structural measures 138
J. Schanze, G. Hutter, E. Penning-Rowsell, D. Parker, H.-P. Nachtnebel, C. Neuhold, V. Meyer & P. Königer
Long term planning, integrated portfolios & spatial planning
The OpenMI-LIFE project – putting integrated modelling into
practice in flood management 141
D. Fortune
A method for developing long-term strategies for flood risk management 142 K.M. de Bruijn, M.J.P. Mens & F. Klijn
Flood Risk Mapping, using spatially based Systems Engineering 143
R. Raaijmakers
Finding a long term solution to flooding in Oxford: The challenges faced 144 L.G.A. Ball, M.J. Clegg, L. Lewis & G. Bell
Risk analysis and decision-making for optimal flood protection level in urban
river management 145
M. Morita
An integrated risk-based multi criteria decision-support system for flood protection
measures in riversheds—REISE 146
N.P. Huber, D. Bachmann, H. Schüttrumpf, J. Köngeter, U. Petry, M. Pahlow, A.H. Schumann, J. Bless, G. Lennartz, O. Arránz-Becker, M. Romich & J. Fries
Integrated methodologies for flood risk management practice
in European pilot sites 148
J. Schanze, P. Bakonyi, M. Borga, B. Gouldby, M. Marchand, J.A. Jiménez & H. Sterr
Underpinning flood risk management: A digital terrain model for the 21st century 150 M. Stileman & D. Henderson
Integrated land and water management in floodplains in England 151
SUDS planning support tool 154 V.R. Stovin, S.L. Moore, S.H. Doncaster & B. Morrow
Strategic planning for long-term Flood Risk Management – findings from case studies
in Dresden and London 155
G. Hutter & L. McFadden
Extreme flood events & flood management strategy at the Slovak-Austrian part of the
Morava river basin 156
M. Lukac & K. Holubova
Using non-structural responses to better manage flood risk in Glasgow 157 R. Ashley, R. Newman, F. McTaggart, S. Gillon, A. Cashman, G. Martin & S. Molyneux-Hodgson
Vulnerability and resilience, human and social impacts
The policy preferences of citizens, scientists and policy makers 161
J.H. Slinger, M. Cuppen & M. Marchand
Analysis of the human and social impacts of flooding in Carlisle 2005 and Hull 2007 162 P. Hendy
Institutional and social responses to flooding from a resilience perspective 163 N. Watson, E. Kashefi, W. Medd, G. Walker, S. Tapsell & C. Twigger-Ross
Flood, vulnerability and resilience: A real-time study of local recovery following the floods
of June 2007 in Hull 164
R. Sims, W. Medd, E. Kashefi, M. Mort, N. Watson, G. Walker & C. Twigger-Ross
Increasing resilience to storm surge flooding: Risks, social networks and local champions 165 H. Deeming
A new model to estimate risk to life for European flood events 166
S.M. Tapsell, S.J. Priest, T. Wilson, C. Viavattene & E.C. Penning-Rowsell
Towards flood risk management with the people at risk: From scientific analysis to practice
recommendations (and back) 167
A. Steinf ührer, C. Kuhlicke, B. De Marchi, A. Scolobig, S. Tapsell & S. Tunstall
Use of human dimensions factors in the United States and European Union 168 S. Durden & C.M. Dunning
Double whammy? Are the most at risk the least aware? A study of environmental justice
and awareness of flood risk in England and Wales 169
J.L. Fielding
Improving public safety in the United States – from Federal protection to shared
flood risk reduction 170
E.J. Hecker, L.J. Zepp & J.R. Olsen
Evaluating the benefits and limitations of property based flood resistance
and resilience – a UK perspective 171
XII
Overcoming the barriers to household-level adaptation to flood risk 173 T. Harries
Human vulnerability to flash floods: Addressing physical exposure and behavioural questions 174 I. Ruin, J.-D. Creutin, S. Anquetin, E. Gruntfest & C. Lutoff
Assessment of extremes
Estimating extremes in a flood risk context. The FLOODsite approach 177 A. Sanchez-Arcilla, D. Gonzalez-Marco & P. Prinos
Inter-site dependence in extremes: Unlocking extra information 178
D.W. Reed
The Flood Estimation Handbook and UK practice: Past, present and future 179 E.J. Stewart, T.R. Kjeldsen, D.A. Jones & D.G. Morris
Extreme precipitation mapping for flood risk assessment in ungauged basins
of the upper Hron River basin in Slovakia 180
S. Kohnová, J. Szolgay, K. Hlavcˇová, L. Gaál & J. Parajka
River flood frequency approaches for ungauged sites 181
A. Calver & E.J. Stewart
Non-stationary point process models for extreme storm surges 182
P. Galiatsatou & P. Prinos
Bayesian non-parametric quantile regression using splines for modelling wave heights 183 P. Thompson, D. Reeve, J. Stander, Y. Cai & R. Moyeed
Multiscale probabilistic risk assessment 185
C. Keef, R. Lamb, P. Dunning & J.A. Tawn
Improving the understanding of the risk from groundwater flooding in the UK 186 D.M.J. Macdonald, J.P. Bloomfield, A.G. Hughes, A.M. MacDonald, B. Adams &
A.A. McKenzie
Radar observation of storm rainfall for flash-flood forecasting 187
G. Delrieu, A. Berne, M. Borga, B. Boudevillain, B. Chapon, P.-E. Kirstetter, J. Nicol, D. Norbiato & R. Uijlenhoet
Climate change impact on hydrological extremes along rivers in Belgium 189 O.F. Boukhris & P. Willems
Uncertainties in 1D flood level modeling: Stochastic analysis of upstream discharge
and friction parameter influence 190
N. Goutal, P. Bernardara, E. de Rocquigny & A. Arnaud
Civil contingency, emergency planning, flood event management
Reservoir safety in England and Wales – reducing risk, safeguarding people 193 I.M. Hope & A.K. Hughes
A comparison of evacuation models for flood event management – application
on the Schelde and Thames Estuaries 194
M.J.P. Mens, M. van der Vat & D. Lumbroso
Hydrodynamic and loss of life modelling for the 1953 Canvey Island flood 195 M. Di Mauro & D. Lumbroso
R. Cossu, Ph. Bally, O. Colin, E. Schoepfer & G. Trianni
Benefits of 2D modelling approach for urban flood management 199
E. David, M. Erlich & A. Masson
Computer modelling of hydrodynamic conditions on the Lower Kuban under various scenarios and definition of limiting values of releases from the Krasnodar, Shapsugsky and Varnavinsky
hydrounits for prevention of flooding 200
M.A. Volinov, A.L. Buber, M.V. Troshina, A.M. Zeiliguer & O.S. Ermolaeva
Flood warning in the UK: Shifting the focus 202
C.L. Twigger-Ross, A. Fernandez-Bilbao, G.P. Walker, H. Deeming, E. Kasheri, N. Watson & S. Tapsell
New approaches to ex-post evaluation of risk reduction measures: The example of flood
proofing in Dresden, Germany 203
A. Olfert & J. Schanze
Dilemmas in land use planning in flood prone areas 204
A. Scolobig & B. De Marchi
Emergency management of flood events in Alpine catchments 205
H. Romang & C. Wilhelm
Flood forecasting and warning
Flood warning in smaller catchments 209
H. Romang, F. Dufour, M. Gerber, J. Rhyner, M. Zappa, N. Hilker & C. Hegg
A prototype of road warning system in flood prone area 210
P.-A. Versini, E. Gaume & H. Andrieu
Snow and glacier melt – a distributed energy balance model within a flood forecasting system 211 J. Asztalos, R. Kirnbauer, H. Escher-Vetter & L. Braun
Analysis of weather radar and rain gauges for flood forecasting 212
M.T.J. Bray, D. Han, I. Cluckie & M. Rico-Ramirez
Integration of hydrological information and knowledge management for rapid decision-making
within European flood warning centres 213
F. Schlaeger, D. Witham & R. Funke
Local warning systems in Slovakia 215
D. Lešková, D. Kyselová, P. Roncˇák & M. Hollá
The provision of site specific flood warnings using wireless sensor networks 216 P. Smith, K. Beven, W. Tych, D. Hughes, G. Coulson & G. Blair
Managing flood risk in Bristol, UK – a fluvial & tidal combined forecasting challenge 217 M. Dale, O. Pollard, K. Tatem & A. Barnes
Off-line flood warning concept for railways 218
U. Drabek, T. Nester & R. Kirnbauer
XIV
Potential warning services for groundwater and pluvial flooding 220
D. Cobby, R. Falconer, G. Forbes, P. Smyth, N. Widgery, G. Astle, J. Dent & B. Golding Data assimilation and adaptive real-time forecasting of water levels
in the river Eden catchment, UK 221
D. Leedal, K. Beven, P. Young & R. Romanowicz
To which extent do rainfall estimation uncertainties limit the accuracy of flash flood forecasts? 222 L. Moulin, E. Gaume & Ch. Obled
Advances in radar-based flood warning systems. The EHIMI system and the experience
in the Besòs flash-flood pilot basin 223
C. Corral, D. Velasco, D. Forcadell, D. Sempere-Torres & E. Velasco
Flash flood risk management: Advances in hydrological forecasting and warning 224 M. Borga, J.-D. Creutin, E. Gaume, M. Martina, E. Todini & J. Thielen
Decision support system for flood forecasting in the Guadalquivir river basin 225 L. Rein, A. Linares, E. García & A. Andrés
Operational flash flood forecasting chain using hydrological
and pluviometric precursors 226
G. Brigandì & G.T. Aronica
Online updating procedures for flood forecasting with a continuous rainfall-runoff-model 227 B. Kahl & H.P. Nachtnebel
GIS technology in water resources parameter extraction in flood forecasting 228 V. Ramani Bai, G. Ramadas & R. Simons
Combining weather radar and raingauge data for hydrologic applications 230 C. Mazzetti & E. Todini
The worst North Sea storm surge for 50 years: Performance of the forecasting system
and implications for decision makers 231
K.J. Horsburgh, J. Williams, J. Flowerdew, K. Mylne & S. Wortley
Probabilistic coastal flood forecasting 232
P.J. Hawkes, N.P. Tozer, A. Scott, J. Flowerdew, K. Mylne & K. Horsburgh
Coastal flood inundation modelling for North Sea lowlands 234
S. Burg, F. Thorenz & H. Blum
New north east of England tidal flood forecasting system 235
A. Lane, K. Hu, T.S. Hedges & M.T. Reis
Impact of extreme waves and water levels in the south Baltic Sea 236
H. Hanson & M. Larson
Bayesian rainfall thresholds for flash flood guidance 238
M.L.V. Martina & E. Todini
Environmental impacts, morphology & sediments
Assessment of hydraulic, economic and ecological impacts of flood polder
management – a case study from the Elbe River, Germany 241
S. Förster & A. Bronstert
Development of estuary morphology models 242
J.M. Huthnance, A. Lane, H. Karunarathna, A.J. Manning, D.E. Reeve, P.A. Norton, A.P. Wright, R.L. Soulsby, J. Spearman, I.H. Townend & S. Surendran
Predicting beach morphology as part of flood risk assessment 246 J.M. Horrillo-Caraballo & D.E. Reeve
Alkborough scheme reduces extreme water levels in the Humber Estuary and
creates new habitat 248
D. Wheeler, S. Tan, N. Pontee & J. Pygott
Managing coastal change: Walberswick to Dunwich 249
M. Cali, A. Parsons, N. Pontee, L. Batty, S. Duggan & P. Miller
Uncertainties in the parameterisation of rainfall-runoff-models to quantify land-use effects
in flood risk assessment 250
A. Wahren, K.H. Feger, H. Frenzel & K. Schwärzel
Impact of the barrage construction on the hydrodynamic process in the severn estuary using
a 2D finite volume model 253
J. Xia, R.A. Falconer & B. Lin
Risk sharing, equity and social justice
From knowledge management to prevention strategies: The example of the tools developed
by French insurers 257
J. Chemitte & R. Nussbaum
What’s ‘fair’ about flood and coastal erosion risk management? A case study evaluation
of policies and attitudes in England 258
C. Johnson, S. Tunstall, S. Priest, S. McCarthy & E. Penning-Rowsell
Flood risk perceptions in the Dutch province of Zeeland: Does the public still support
current policies? 259
J. Krywkow, T. Filatova & A. van der Veen
A partnership approach – public flood risk management and private insurance 260 M. Crossman, S. Surminski, A. Philp & D. Skerten
The international teaching module FLOODmaster – an integrated part of a European educational
platform on flood risk management 262
J. Seegert, C. Bernhofer, K. Siemens & J. Schanze
Decision support for strategic flood risk planning – a generic conceptual model 263 A.G.J. Dale & M.V.T. Roberts
Who benefits from flood management policies? 264
N. Walmsley, E. Penning-Rowsell, J. Chatterton & K. Hardy
Uncertainty
Long term planning – robust strategic decision making in the face of gross uncertainty
(tools and application to the Thames) 267
C. Mc Gahey & P.B. Sayers
XVI
Staged uncertainty and sensitivity analysis within flood risk analysis 269 B. Gouldby & G. Kingston
Assessing uncertainty in rainfall-runoff models: Application of data-driven models 270 D.L. Shrestha & D.P. Solomatine
Flash floods
European flash floods data collation and analysis 273
V. Bain, O. Newinger, E. Gaume, P. Bernardara, M. Barbuc, A. Bateman, J. Garcia, V. Medina, D. Sempere-Torres, D. Velasco, L. Blaškovicˇová, G. Blöschl, A. Viglione, M. Borga, A. Dumitrescu, A. Irimescu, G. Stancalie, S. Kohnova, J. Szolgay, A. Koutroulis, I. Tsanis, L. Marchi & E. Preciso
Representative flash flood events in Romania Case studies 275
G. Stancalie, B. Antonescu, C. Oprea, A. Irimescu, S. Catana, A. Dumitrescu, M. Barbuc & S. Matreata
Changes in flooding pattern after dam construction in Zadorra river (Spain): The events
of October 1953 and February 2003 276
A. Ibisate
Post flash flood field investigations and analyses: Proposal of a methodology and illustrations
of its application 277
E. Gaume & M. Borga
Hydrological and hydraulic analysis of the flash flood event on 25 October 2007 in
North-Eastern part of Sicily, Italy 278
G.T. Aronica, G. Brigandì, C. Marletta & B. Manfrè
The day roads became rivers: A GIS-based assessment of flash floods in Worcester 279 F. Visser
Risk and economic assessments
Flood risk mapping of Austrian railway lines 283
A. Schöbel, A.H. Thieken & R. Merz
Correlation in time and space: Economic assessment of flood risk with the Risk Management
Solutions (RMS) UK River Flood Model 284
D. Lohmann, S. Eppert, A. Hilberts, C. Honegger & A. Steward-Menteth
A case study of the Thames Gateway: Flood risk, planning policy and insurance loss potential 285 J. Eldridge & D.P. Horn
Integration of accurate 2D inundation modelling, vector land use database and economic
damage evaluation 286
J. Ernst, B.J. Dewals, P. Archambeau, S. Detrembleur, S. Erpicum & M. Pirotton
Planning for flood damages reduction: A case study 287
M. Karamouz, A. Moridi & A. Ahmadi
High resolution inundation modelling as part of a multi-hazard loss modelling tool 288 S. Reese & G. Smart
Estimation of flood losses due to business interruption 289
I. Seifert, H. Kreibich, B. Merz & A. Thieken
Residential flood losses in Perth, Western Australia 290
Development of a damage and casualties tool for river floods in northern Thailand 293 J.K. Leenders, J. Wagemaker, A. Roelevink, T.H.M. Rientjes & G. Parodi
Synthetic water level building damage relationships for GIS-supported flood vulnerability
modeling of residential properties 294
M. Neubert, T. Naumann & C. Deilmann
Impacts of the summer 2007 floods on agriculture in England 295
H. Posthumus, J. Morris, T.M. Hess, P. Trawick, D. Neville, E. Phillips & M. Wysoki
Climate change
Simulating flood-peak probability in the Rhine basin and the effect of climate change 299 A.H. te Linde & J.C.J.H. Aerts
Climate changes in extreme precipitation events in the Elbe catchment of Saxony 300 C. Görner, J. Franke, C. Bernhofer & O. Hellmuth
A methodology for adapting local drainage to climate change 301
R.M. Ashley, J.R. Blanksby, A. Cashman & R. Newman
Exploring and evaluating futures of riverine flood risk systems – the example of the Elbe River 303 J. Luther & J. Schanze
Foreword
Since the dawn of civilisation human society has been shaped by its interaction with water—whether too much or too little. Indeed, water is one of the powerful forces of nature that have formed the planet on which we live. Floods are not new—they continually make news because of both tragic effects on individuals and acts of heroism in the emergency. The impact of floods on people can be dramatic even if there are no fatalities; pictures of rescue make good television material with stories told of bravery and fortunate escapes. However, the aftermath of a flood is distressing with personal possessions ruined, houses deep in sewage-contaminated mud, vital services disrupted and businesses destroyed. Cleaning up, community recovery and economic restoration takes months and for many there remains the fear that the next time it rains or a severe storm is predicted, the experience will be repeated.
When discussing flood risk it is important to remember that “risk” is entirely a human concern; floods from river, estuary or coast are predominantly natural events; they are random. The risk arises because the human use and value of the river and coastal plains conflicts with their natural functions of storage and movement of water during a flood. Of course, the potential causes for some flooding are man-made, for example following the breaching of a dam or a flood embankment, and some floods are triggered by other hazards such as tsunami fol-lowing an earthquake. In many respects the types of impact on people are similar to the more common sources of flooding—but probably they are more severe as these are often less predictable events. Catalogues of recent disasters are common in the introduction to volumes such as this, but we do not dwell here on recent floods, or on the effect of climate change, as there is much in the text.
It is, however, essential to comment on a major development in flood management policy from the European Union which will affect flood risk management in all EU Member States. On the 26th November 2007 the European Directive on the assessment and management of flood risks was enacted. This will be transposed into national legislation in each Member State within 2 years and sets out a set of actions on preliminary flood risk assessment, flood risk mapping and the preparation of flood risk management plans to be completed by the end of 2015. The Directive covers all sources of flooding (not just rivers, but coastal floods, urban and groundwater floods); it requires planning at basin scale and has specific requirements for trans-national basins; and, in all cases the potential impacts of climate change on the flood conditions need to be considered. The use of the phrase “management of flood risks” in the title of the Directive indicates that European policy has progressed away from a philosophy of flood control to the acceptance that flood risks should be managed.
FLOODrisk 2008 marks the completion of some substantial research projects: • FLOODsite—an Integrated Project in the EC Sixth Framework Programme • The first phase of the Flood Risk Management Research Consortium • The first common call of the CRUE ERA-NET
FLOODsite is the largest ever EC research project on floods, and will be completed in early 2009. The FLOOD-site consortium involves 37 of Europe’s leading institutes and universities and brings together scientists from many disciplines along with public and private sector involvement from 13 countries. There are over 30 project tasks including the pilot applications in Belgium, the Czech Republic, France, Germany, Hungary, Italy, the Netherlands, Spain, and the UK. FLOODsite covers the physical, environmental, ecological and socio-economic aspects of floods from rivers, estuaries and the sea. In this volume there are papers on many aspects of the FLOODsite project.
The Flood Risk Management Research Consortium (FRMRC) was established in the UK by the Engineering and Physical Sciences Research Council to undertake an integrated programme of research to support effective flood risk management by
XX
• short-term delivery of tools and techniques to support short term improvements in flood risk management in the United Kingdom; and,
• development and training of the next generation of flood risk management professionals through their involvement in and exposure to the consortium’s research.
The FRMRC involved over 20 UK universities with research in eight priority areas. FRMRC has recently completed the first phase of its multi-disciplinary programme of research; this volume contains several papers on their results. A second phase of the FRMRC programme commenced during 2008.
Whereas FRMRC and FLOODsite are both research projects, the CRUE ERA-NET does not directly carry out research. Rather it is a network of the major research funders in the EU who are exploring how to integrate their national research programmes more closely as part of the EU policy to strengthen the European Research Area. As part of this closer cooperation the CRUE partners have developed a vision for the future research needed and issued a common call for research on non-structural measures for flood risk management. The concept of the common call was to explore how national programmes with their different regulations could work together in identifying research topics and jointly tendering, commissioning, monitoring and evaluat-ing research projects. The scientific advances from these first common call projects are presented within this volume.
In setting up FLOODrisk 2008 our intention was to cover flood risk management in an integrated and com-prehensive way. Thus the call and selection of papers covered the physical and social sciences, included policy and practice, and ranged from long-term planning, emergency management and post-flood recovery. The theme of the conference is research into practice and we hope that much of the research discussed at FLOODrisk 2008 will improve the scientific evidence and practice in the actions of the Floods Directive in Europe and find application worldwide.
It is our pleasure to welcome you to the FLOODrisk 2008 conference.
Stephen Huntington
Chairman of the International Scientific Committee
Paul Samuels
Committees
INTERNATIONAL SCIENTIFIC COMMITTEE
Professor Stephen Huntington (Chair) HR Wallingford, UK
Dr Peter Bakonyi VITUKI, Budapest, Hungary
Professor Eelco van Beek Deltares, The Netherlands Professor Marco Borga Università di Padova, Italy
Professor Ian Cluckie University of Bristol, UK
Professor Jean-Dominique Creutin INP Grenoble, France
Dr Rolf Deigaard DHI Group, Denmark
Ronnie Falconer Jacobs / EWA, UK
Dr Frans Klijn Deltares, The Netherlands
Dr Andreas Kortenhaus Universität Braunschweig, Germany
Dr Elisabeth Lipiatou DG Research, Brussels, EU
Dr Joan Pope US Army Corps of Engineers, USA
Professor Panos Prinos Aristotle University, Thessaloniki, Greece
Professor Paul Samuels HR Wallingford, UK
Paul Sayers HR Wallingford, UK
Professor Agustin Sanchez-Arcilla Universitat Politècnica de Catalunya, Spain
Dr Patrick Sauvaget SOGREAH, France
Dr Jochen Schanze IÖR, Dresden, Germany
Professor Gheorge Stancalie NMA, Bucharest, Romania
LOCAL ORGANISING COMMITTEE
Professor Paul Samuels HR Wallingford, UK
Professor Garry Pender Heriot-Watt University, UK
Professor William Allsop HR Wallingford, UK
Jackie Harrop HR Wallingford, UK
Chris Grandy Creative Conferences, UK
Coastal flooding: A view from a practical Dutchman on present
and future strategies
J.W. van der Meer
Van der Meer Consulting BV, Heerenveen, The Netherlands
ABSTRACT: This key note paper intends to feed further discussion on safety against coastal flooding. It will mainly be based on the Dutch situation, where half of the population lives below sea level, but the paper will give enough discussion points for other situations. Observations, conclusions, etc., made in this paper and the presentation are on the personal account of the author, so do not represent any official view from the Netherlands.
The paper briefly describes the history of creating safety against flooding, which started after the large flood-ing in 1953 in the south west of the Netherlands with almost 2000 casualties. This led to the situation with high and strong dikes, which should withstand a storm with a certain return period between 2,500 and 10,000 years. A discussion started already 15 years ago on how to derive new rules, based on probability of flooding or even on flood risk. This discussion continues, but is now fed with many more calculations on failure of flood defence assets, breaching, inundation, damage, evacuation and last but not least: indestructible dikes.
1 INTRODUCTION
Coastal flooding has always been an important issue in the Netherlands, mainly because the whole country covers more or less the delta of the rivers Rhine and Meuse, and river delta’s are by definition low com-pared to the sea. By protecting the low lying areas with dikes, the areas themselves settled by a metre or more and became even lower than the natural delta. Protection against flooding became more and more important.
The driving force for coastal flooding in the Netherlands will always be a very severe storm. In other countries also hurricanes or tsunamis may be driving forces. River flooding in the Netherlands, however, is closely linked to coastal flooding, mainly for two reasons. First there are estuarine areas where both a storm or a high river discharge may give flood-ing. The second reason is that the whole safety system in the Netherlands is not separated in coastal or river flooding, but is simply based on flooding in general. The paper will discuss some items where this may lead to wrong interpretations, basically due to not understanding fully the difference between the two driving forces, severe storm or high river discharge.
The word “dike” means any structure made out of soil (sand, clay), often protected by a kind of
revet-countries may use terms like levees or embankments, but the structures are more or less similar.
The paper will cover past, present and future strat-egies. Interest in coastal flooding is increasing in the Netherlands and not only by coastal or civil engineers. Recently, this has widened the scope of feasibility studies to explore all kind of ideas, like insurance, evacuation, awareness, compartment (dividing a flood risk area in two parts, reducing the consequences of flooding) and also indestructible dikes. This last option would mean a flooding probability of (almost) zero and therefore a flood risk of almost zero.
As already noted, observations, conclusions, etc., made in this paper and the presentation are on the per-sonal account of the author, so do not represent any official view from the Netherlands.
2 DECISIONS AFTER THE 1953 FLOOD Early February, 1953, a severe storm hit the south west part of the Netherlands and also parts of Belgium and the UK, causing severe flooding with, in the Nether-lands, almost 2000 casualties. Although people had warned before about the fairly low dikes and the real possibility of a major flood, the interest in those days after the world war was more directed to build up
4 After the flood the Delta Committee was formed with the main goal to present a safety policy against flooding for the future. In those days they performed a kind of flood risk analysis. They concluded that the probability of flooding for central Holland (Amsterdam, The Hague, Rotterdam) should be around 1/125,000 per year. But they wanted or had to be practical, and they understood that calculating probability of failure, including all failure mecha-nisms of dikes, was not yet possible.
The outcome was: design a safe dike for an event with a probability of 1/10,000 per year. This had two advantages. First it was clear for what kind of event the dikes should be designed and secondly, normal design procedures could be used (instead of describ-ing failure mechanisms leaddescrib-ing to flooddescrib-ing, which is necessary for flood risk design).
But the principal of flood risk was not forgotten. It was clear that some parts of the country had more inhabitants and more investments than other parts and consequences of flooding, therefore, would be different. Each part got his own “event” to design for: 1/10,000 per year for central Holland, 1/4,000 per year for most others and for smaller areas even 1/2,000 per year.
Later on, also the rivers were included in the safety policy. It was realized that evacuation would be pos-sible for flooding from a high river discharge, as it would be predicted some days before. This would lead to less casualties and, therefore, most river dikes had to be designed for a water level which would have a probability to occur of 1/1,250 per year. Since then the safety against flooding always considers both, coastal flooding and river flooding. Figure 1 shows all primary flood defences.
It was Edelman (1954) who realized that if three weak points were present at a dike section under severe wave attack, it would fail:
1. If the crest was too low, this would lead to exten-sive overtopping;
2. If bad quality of material was present, infiltration of water in the dike would be fast;
3. If a steep inner slope was present, it would lead to a slip failure when wet.
Based on analysis of dike breaches in 1953, Edelman concluded that if one of these 3 items was not present, then very often there was no breach. His suggestion was to make inner slopes much more gen-tle, like 1:3, but allow overtopping. He was convinced that a dike could withstand wave overtopping, as long as the inner slope would be gentle enough.
The final decision for design, however, was differ-ent. It was indeed decided to make gentler inner slopes of 1:3, but moreover, not to allow (severe) wave over-topping. The crest height should be designed equal to
the 2%-wave run-up level. It was expected that any dike crest and inner slope with grass cover would resist 2% of the incoming waves overtopping the crest.
With this design principle all sea dikes have been improved since 1953 and actually, present designs still use these principles. In the nineties the 2%-wave run-up level changed to 1 l/s per m wave overtopping.
3 SAFETY ASSESSMENT
After improvement of most of the dikes in the Neth-erlands and construction of the storm surge barriers in the Eastern Scheldt and the entrance to the port of Rotterdam, it was realized that the flood protection system should not only be designed and constructed, but should also regularly be checked. The Flood Defence Act of 1996 ruled that every 5 years a safety assessment should be performed on all primary flood defence assets.
This safety assessment has been based on the same principle as for the design: the dike or flood protec-tion asset should be safe for a certain event with a cer-tain probability of occurrence. But there are cercer-tainly differences between safety assessment and design.
In a design the actual properties of the material of the dike are not known, but assumed. Safety fac-tors are taken into account and a little more safety does not cost a lot more as it will all be part of a new or improved structure. In a safety assessment the structure is present and material properties can be measured. But including more safety means that
Figure 1. The Netherlands as delta with all primary flood defences, both for coastal and river protection.
breach”.
Reality is different. Experience shows that where doubt is present, the dike section will be disquali-fied. There is probably another reason behind these decisions, not stated publicly. The water boards have to maintain the majority of the dikes. They have to pay the maintenance from local taxes they earn. But major improvements, as a consequence of the safety assessment, will be paid by the government. It is for this reason that water boards can not completely be objective in the safety assessment procedure. There is benefit in obtaining an improved protection.
The safety assessment is quite complex as it has to consider all parts of a dike or flood defence asset, for all kind of failure mechanisms. Certainly in the first assessments, parts were discovered which did not pass the assessment criteria or where assessment criteria were not yet available. In the latter case also design rules were not available and actual design had always been based on experience rather than design rules.
When the results of the first assessments were summarized, it appeared that in about one-third of all the dike sections, parts were disqualified (and had or have to be improved) or an assessment rule was not available (and therefore no assessment result was available). This has been interpreted from two sides. One side states that even with disqualification of parts, there is not a direct threat for flooding and there will be sufficient time to design and improve the part of the dike, such as a stronger revetment or a little higher crest. Lacking knowledge means that this knowledge has to be developed. The other side states that only two-third of all flood defence assets are safe and that the other one-third gives a serious threat. So politicians should release more money for improving dikes and developing knowledge.
A more general conclusion is that the Netherlands has never been more safe against flooding than in the present situation, but that still quite some work has to be done to be safe in agreement with the safety assessment rules.
4 FROM PROBABILITY OF EVENT TO FLOODRISK
The present design and safety assessment rule is that the flood defence asset should withstand an event with a certain return probability or probability per year. It
and contra’s. The probability of flooding is easy to explain to the public: it gives the probability per year that it is expected to get wet feet. The difficulty is that one needs a full description of each failure mode from initial damage up to the initiation of a breach.
Work under Task 4 of FloodSite (Allsop et al., 2007) made a good step by describing most of the failure modes. But the result of a calculation is never better than the failure mode modeled. VNK 1 was the first attempt to calculate flooding probabilities for various areas in the Netherlands. (VNK stands for Safety of the Netherlands calculated and mapped). Real dike ring areas were considered and all possible failure mechanisms. It took a few years by a number of consortia to come up with results for some 10–15 dike ring areas (where we have more than 50).
One conclusion or result was that by this proce-dure it is easy to find the weakest locations in a dike ring and for what failure mechanism. Upgrading that section directly reduces the probability of flooding. It might be noted, however, that these “weak” sec-tions can also be found by applying the regular safety assessment. The probabilistic method, however, gives how much the probability of flooding would improve, which is not possible with the safety assessment.
The calculations by VNK 1 also showed that some failure mechanisms were not well understood or not modeled well enough. And in such a case uncertainty is taken into account which sometimes led to unreal-istically large flooding probabilities. In such a case more study is required to improve the modeling of the failure mechanisms.
Since VNK 1 the modeling has improved and pro-duction runs will be made in 2008/2009 to calculate flooding probabilities of all 53 dike ring areas in the Netherlands under VNK 2.
During (and after) VNK 1, a lot more information became available on the consequences of flooding. Numerical tools were developed to model realistically the water flow and inundation in time, assuming one or more initial breaches in the dike ring system. Dam-ages were calculated as well as casualties. Extreme assumptions were made to find upper boundaries. Moreover, it gave insight in inundation depths and the most vulnerable locations in the Netherlands.
Flood risk is the product of probability of flooding and consequences, so probability multiplied by cost (money). There is similarity with an insurance pre-mium. A flood risk could be for example 2 million
6 goal seemed to be to come to regulations based on true flood risk, nowadays the insight has changed a little. Probability of flooding, as calculated by VNK 2, will probably be taken as the primary result. In future flood probability may become the normative rule. The insight in consequences (damage, casualties) will steer the normative rule, not the product of probabil-ity and damage.
5 SAFETY UP TO 2100
Many feasibility studies are going on in the Neth-erlands and safety against flooding now has interest from a wider professional audience than just civil engineers. On 19 June 2008 a one day conference/ workshop (general presentation, workshop discus-sions, no papers available) was held with the title “The power of water”. This conference released the policy for flood defence in the Netherlands and con-sisted of 3 layers:
1. Prevention is and stays number one. It is always better to prevent anything to happen than to mini-mize the consequences. More knowledge should be gained on the actual strength of flood defence assets, consequences should be studied and cli-mate change should be taken into account. Innova-tive solutions should be studied, like indestructible dikes.
2. Spatial planning should include safety against flooding.
3. Reduce remaining consequences by evacuation and awareness.
All points will be elaborated a little more, starting with the new points 2 and 3. Policy makers believe that spatial planning can work if a safety assessment procedure will be part of it. It should lead to decisions not to built new houses or industry in some parts, where for example the area is many meters below sea level. Or it should lead to decisions to raise the level several meters before starting construction.
A large part of the conference did not believe that this second layer would work. The main reason is that spatial planning is in the hands of the local authorities who decide on it, not the government. Local authori-ties will always decide to improve their own area and will never say: go to the town 10 km further, because their level is higher than here! Another reason is that flooding by sea or river is not an issue in daily politics of a local area.
Evacuation belongs to the third layer. Due to the fact that discussion on flooding always includes both sea and river flooding, some interpretations of phe-nomena are considered true in both situations. For evacuation this is certainly not the case and only a few people are aware of it.
A high river discharge in the Netherlands, with consequently a high water level against the dikes, is not caused by flash floods, but by very heavy rain in Switzerland and the south of Germany, and prob-ably the Netherlands. It takes days before this water comes to the border of the Netherlands and good computer models are available to predict where, when and how high the water level will come along the river dikes.
In case of an emergency, where predicted water lev-els may indicate an unsafe situation, there is time to evacuate thousands of people. In 1995, 100,000 peo-ple were evacuated in a situation where some dikes were not yet improved and where the safety could not be guaranteed during that high water. In those cases the weather is not too bad for evacuation and there is time enough.
Also for a hurricane, like in the US, there is time to evacuate. Evacuation in such a situation, however, is mainly based on the destructive wind along the coast, not entirely on a probability of flooding.
The possibility of evacuation is often transferred to coastal situations. And this is a complete mistake, cer-tainly for Dutch situations! It may be possible in the UK in rural areas, where for example a small number of farmers live in a relatively small flood risk area, protected by dikes only able to protect against events smaller than 1/30 or 1/100 years (or even less). For each very severe storm warning they should evacuate. But here it will be a very small number of people who are aware of the situation.
Assuming that the dikes in the Netherlands can withstand an event with a return period of 10,000 years, evacuation would only be an option if a storm is expected which would even be worse. At present we are not able to predict whether a storm would be an event with a smaller or larger return period than 10,000 years. It depends also on the local conditions like tide where the worst condition with respect to maximum surge level and wave conditions will occur. This means that we should evacuate the whole north, west and south west of the Netherlands, say around 5–10 million people, for a storm warning with a return period in the order of 1,000 years or more.
But would that be possible? Such a severe storm will already have a wind force close to Beaufort 11 one day before the actual peak of the storm (with then certainly Beaufort 12 and more). Such a 1/10,000 years storm will have a devastating effect on the country. Many roofs will blow away, thousands of trees will break, tiles and everything which was not tied thoroughly, will fly through the air.
In 1999 a short but very strong storm hit the Atlan-tic coast north of Bordeaux in France. Large areas with trees were completely destroyed. Even after more than two years the trees were not yet all removed, see Figure 2. The storm led to flooding in the Gironde.
It may be clear: nobody wants to evacuate in such a storm. It would be very dangerous. The only option is to wait in a safe place and, indeed, if a flooding occurs, go to the first or second floor and hope that the house will be strong enough to withstand the water.
So in planning any such evacuation, the division between coastal flooding and river flooding must be made, and it may be wiser not to assume that evacua-tion is always possible.
It is, however, always good to increase awareness for disasters, like a flooding, which is the second item of the third layer (evacuation and awareness). Dur-ing such an event power may be shut down, as well as energy, water supply, etc. Awareness and prepara-tions are good, not only for a disaster like flooding, but actually for all possible disasters.
6 SAFETY OF COASTAL DIKES 6.1 Main failure mechanisms
Coastal dikes are designed for high storm surges and related severe wave attack. Both the high water level and the waves give the loading to the dike. Two main failure modes exist for coastal dikes. One is the fail-ure of the seaward protection by large waves. A many small and large scale model tests have been performed in wave flumes to find the relationship between wave attack and strength of a variety of revetments, from rock revetments to asphalt layers. We can conclude that we know a lot on strength of these kind of protec-tion systems.
The other main failure mechanism is wave over-topping and failure of the inner slope of the dike. We know a lot about wave overtopping, or actually, the
indeed some tests have been performed in these facil-ities, in the past and recently.
The fact that the hydraulic behaviour of wave over-topping is known, has led to the idea of the Wave Overtopping Simulator. This new device has been used for erosion tests performed on several real dikes and insight in strength has gained tremendously. Results will be summarized here.
6.2 Erosion by wave overtopping
Two mechanisms may lead to failure due to wave overtopping. The first is infiltration of overtopping water into the dike and eventually sliding of the inner slope. The second is erosion of the cover layer of clay and grass by overtopping waves, followed by erosion of the inner slope (clay or clay layer on sand core).
The first mechanism, infiltration and sliding, can only occur if the inner slope is quite steep, see also the points mentioned by Edelman, 1954, in Chapter 2. For this reason most coastal dike designs in the Neth-erlands, after the flood of 1953, got a 1:3 inner slope. It is assumed that such a slope will not slide due to infiltration of water. But if a steeper slope is present, already 1 l/s per m overtopping would be enough to give sufficient infiltration of water.
This means that for steep inner slopes (steeper than 1:3 or may be 1:2.5) the critical overtopping discharge is already 1 l/s per m. For dikes with an inner slope of 1:3 or gentler we assume that infiltration and sliding is not a governing failure mechanism. Only erosion by overtopping remains.
Till a few years ago hardly anything was known about resistance of inner slopes of dikes with grass against wave overtopping. But in the beginning of 2007 and 2008 innovative erosion tests have been per-formed for various dike sections. In 2006 the Wave Overtopping Simulator was constructed, see Van der Meer et al., 2006. The basic idea is that a constant discharge is pumped into a box on top of a dike and then the pumped volume is released from time to time in such a way that it simulates overtopping waves in reality. Figure 3 gives an impression of the working of this wave overtopping simulator.
Tests have been performed for mean overtopping discharges starting at 0.1 l/s per m up to 75 l/s per m. In 2007 3 dike sections have been tested, which are reported by Van der Meer et al., 2007, Akkerman et al., 2007 and in the ComCoast reports (www.
Figure 2. Two and a half years after a short but devastating storm, all fallen trees have not yet been removed. Gironde, France, June 2002.
8
It seems unlikely that an inner slope with a clay cover topped with a grass cover (in Dutch situations) will fail due to erosion by overtopping waves with a mean dis-charge of 30 l/s per m or less. Future research may result in a final conclusion.
A large number of dike sections withstood 50 l/s per m and some of them even 75 l/s per m. No section failed for 30 l/s per m, which gives the basis for the preliminary conclusion.
7 INDESTRUCTIBLE DIKES 7.1 Case study
The 10–4 event is already very extreme. In stochastic
terms a probability of zero does not exist, but “practi-cally zero” can be defined as: two orders of magnitude more safe than now. If a dike can resist a 1/1,000,000 storm can we give it the title indestructible? What do we have to do to make such a dike?
A short feasibility study was made to explore this idea. Four cases (dike sections) were chosen, one in the north along the Waddensea, one directly on the North Sea coast, one in an estuary and one along the coast of the big lakes. All cases showed for the safety assessment situation (event around 1/10,000 per year) an overtopping discharge around 1 l/s per m.
Wave conditions and water levels were deter-mined for the 10–4, 10–5 and 10–6-events and then
PC-OVERTOPPING was used to calculate the overtopping discharges. These were respectively around 1, 5–10 and 20–30 l/s per m. The 20–30 l/s per m overtopping discharge is still equal to or smaller than the limit of 30 l/s perm.
A preliminary conclusion may be that a design with 1 l/s per m overtopping leads to a robust and “indestructible” dike section (with respect to erosion by overtopping). It should be noted that such a dike should have an inner slope of 1:3 or gentler.
A more extreme event does not only lead to higher water levels, but also to larger waves. Another failure mechanism is stability of the revetment. Most stabil-ity formulae are based on the stabilstabil-ity number Hs/ΔD, where Hs= the significant wave height (at the toe of the dike), Δ = relative mass density and D = a diam-eter or thickness.
A larger wave height leads then linearly to a larger diameter or thickness. The increase in wave height from a 10–4 to a 10–6-event is more or less the same
increase that is required to make the revetment “inde-structible”. In the case study the increase in wave height was 10–25%. The consequence to make an “indestructible” revetment would be to increase the thickness by at least 10–25% and also to apply the revetment protection to a higher level on the dike, as the 10–6 -event has a higher water level.
Figure 3. The Wave Overtopping Simulator releases 22 m3
of water over 4 m width in about 5 s. It simulates a large overtopping wave with a mean discharge of 75 l/s per m.
Figure 4. Damage to a dike section during a test with 75 l/s per m wave overtopping.
Part of the results has been given by Steendam et al., 2008, at this conference. They come to a few preliminary conclusions, mainly based on observa-tion rather than thorough analysis, which still has to be performed. The most important one in relation to actual strength of dikes by wave overtopping is:
have been improved. There is no reason to believe that this tradition will stop. Improvements in the past few decades have always been designed for a life time of 50 years. It can be assumed that in the next 50 years almost all coastal dikes, or at least a major-ity, in the Netherlands will be improved again. That is a unique opportunity to investigate and go for inde-structible dikes.
It is realized that this is perhaps a situation which is only present in the Netherlands. It is different in situ-ations where the present safety is 1/100 per year or less. But even there, prevention is always better than facing a major flood.
7.2 Fragility curves
Safety assessments of flood defence assets are increasingly performed with the technique of structural reliability. All parameters, load param-eters (hydraulic boundary conditions) and strength parameters (dike characteristics), are taken into account and expressed as stochastic variables. One of these structural reliability methods is to calculate the failure probability (Pf) of a flood defence, given a certain water level. Assembling the failure prob-abilities for several water levels constructs a fragil-ity curve, see Van der Meer et al., 2008, presented at this conference.
This paper described the situation in the Nether-lands, where design events have a return period in the order of 10−4 per year. The fragility curve gives the
probability of failure given a certain water level, not a return period of that water level. But in an actual case there is a known relationship between the water level (storm surge), including wave conditions, and the return period of that event. Therefore, it is fairly easy to calculate a fragility curve where the probabil-ity of failure is give as a function of the return period of the water level or event. Figure 5 gives an example for a large sea dike (one of the case studies discussed before).
The graph shows actually three failure modes: 1. Infiltration of overtopping water and sliding of
the inner slope (if the inner slope would be steep). This would occur for an overtopping discharge of 1 l/s per m;
2. Erosion of the inner slope by wave overtopping (the
Piping in this example does not give a serious probability of failure. The 1 l/s per m overtopping discharge gives more or less a probability of failure of 50% for the 10−4-event. This is exactly the design
condition.
But the graph gives also a similar impression as the calculations on indestructible dikes: the 50%-probability for 30 l/s per m in this graph gave a return period of 2.10−6, which is more extreme
than the 10−6-event. One can say that the differences
between the curves for 1 and 30 l/s per m in Figure 5 give the safety between design and failure and that the probabilities for the 30 l/s per m curve actually indicate that this dike section is “indestructible” with respect to erosion by wave overtopping.
8 CONCLUDING REMARKS
The major improvements of coastal dikes in the Netherlands, after the 1953 flood, was based on three principles. Design for an event with a return period around 10,000 years; make inner slopes of a dike at least 1:3; and design for the 2%-run-up level or 1 l/s per m wave overtopping. This has led to high and strong dikes.
A safety assessment procedure was introduced, which has to be performed every 5 years for all flood defence assets. The first assessments showed weak and inadequate parts, which are still being improved.
A new policy on flood defence was released recently, where three layers were introduced. The first still being prevention. The two added layers are to include safety against flooding in spatial planning and to make evacuation plans and to make people more aware of the possibility of a disaster. These two added layers still have to be explored.
The recent destructive tests with the Wave Over-topping Simulator showed that clay with a grass
0.0 0.1 0.2 0.3 0.4
1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08
Return period of water level [years]
Failure probability F
q = 50 l/s/m Design water level
Figure 5. Fragility curves as a function of the return period of the water level.